Validation of Automated Low Voltage Area Measurement Using CARTONET in Atrial Fibrillation Patients

Area of Science:

• Electrophysiology
• Cardiac Mapping
• Artificial Intelligence in Medicine

Background:

• Low-voltage areas (LVAs) are crucial indicators of atrial remodeling and predictors of atrial fibrillation (AF) recurrence post-pulmonary vein isolation (PVI).
• Manual quantification of LVAs is time-consuming, operator-dependent, and subject to variability, hindering consistent clinical application.

Purpose of the Study:

• To validate the accuracy, consistency, and efficiency of CARTONET, an automated cloud-based platform for LVA quantification.
• To compare CARTONET's performance against conventional manual LVA analysis in patients with AF.

Main Methods:

• Retrospective analysis of 100 AF patients undergoing PVI, comparing manual LVA measurements with CARTONET quantification.
• LVA stratification into four categories: low (<5%), mild (5%-20%), moderate (20%-35%), and severe (>35%).
• Statistical validation using Spearman's rank correlation, Bland-Altman agreement, and confusion matrices, with subgroup analysis for segmentation errors and map density.

Main Results:

• CARTONET showed strong correlation with manual measurements (ρ = 0.928-0.983) across all left atrial regions.
• Minimal bias observed in Bland-Altman analysis (mean difference: -0.8%), with 96% of values within agreement limits.
• Categorical agreement for LVA stratification exceeded 90%, with segmentation errors minimally impacting quantification.

Conclusions:

• CARTONET provides reliable and efficient LVA quantification, demonstrating high concordance with manual methods.
• The platform's performance in high-quality electroanatomic mapping supports its clinical and research utility.
• CARTONET reduces operator workload and promotes standardized substrate evaluation for AF management.